Turning Policy into Code using AI

When you spend enough time working in public benefit systems, a familiar frustration starts to surface: the people who understand the policy and the people who build the systems to enforce it often talk past each other.

One group is fluent in regulations, exceptions, and intent. The other speaks in logic, constraints, and system architecture. It’s like putting two brilliant minds in a room with no common language….important things get lost in translation.

I’ve seen that breakdown cause slow, expensive, and sometimes flawed system implementations. and I approached it with one core idea: What if we could close that communication gap—using generative AI?

While many teams focused on improving experiences for applicants or navigators (and rightfully so), I turned my attention upstream—to the policy experts. I wanted to empower the people who best understand the rules to also shape how those rules are implemented in code.

So I started experimenting with large language models to generate a sort of “Rosetta Stone” between policy and software. The result was an intermediate format—a domain-specific language (DSL) built just for describing public benefit policies. It’s structured enough for engineers to implement directly, but still readable and verifiable by policy experts.

This DSL became the connective tissue: a shared understanding between the people who write the rules and the ones who build the systems to carry them out.

I spent the summer working with various LLMs to refine this idea, testing prototypes and testing how reliably models could generate policy logic in this DSL.

This work drew on everything I care about: responsible AI, civic tech, human-centered design, and the promise of government systems that actually work.

If you’re working in the public benefit space—or thinking about how to turn policy into code in your own world—I’d love to share ideas. This is just the beginning.